The Impacts of Natural Gas Prices on the California Economy: Final Report

Transcription

1 The Impacts of Natural Gas Prices on the California Economy: Final Report PREPARED FOR: The California Natural Gas Advisory Group PREPARED BY: Global Insight, Inc. Global Energy Services, United States 24 Hartwell Avenue, Lexington MA Phone: +1 (781) Fax: +1 (781) U.S. Regional Services, United States 800 Baldwin Tower, Eddystone PA Phone: +1 (610) Fax: +1 (610) February 2006

2 TABLE OF CONTENTS EXECUTIVE SUMMARY... 1 INTRODUCTION Study Objectives...12 Requirements of the Study Sponsors...12 BACKGROUND Natural Gas Supply and Demand in California...14 Recent Price Trends...17 STUDY ASSUMPTIONS Natural Gas Price Scenarios...19 Crude Oil and Natural Gas Prices...21 Economic Sectors Dependent on Natural Gas...22 STUDY METHODOLOGY Preparing Energy Forecasts...27 Estimating U.S. Macroeconomic Impacts...28 Estimating California Economic Impacts...28 Revisions to the Global Insight California Forecast Model...28 ECONOMIC IMPACTS Economic Effects of Changes in Natural Gas Prices...31 Impacts on the U.S Economy...31 Impacts on the California Economy...32 Impacts on Major End Users...38 Impacts in Northern and Southern California...43 Results of Other Studies...45 SUMMARY APPENDIX A: DETAILED ECONOMIC IMPACTS IN CALIFORNIA APPENDIX B: INTERVIEWS WITH ENERGY-INTENSIVE INDUSTRIES APPENDIX C: ECONOMIC IMPACTS IN CALIFORNIA BY SECTOR... 65

3 EXECUTIVE SUMMARY Objective The objective of this study is to assess the impacts on the California economy of changes in natural gas prices. Since California is more dependent on natural gas as a source of energy than are most other states, a continuing concern is the negative economic effect that a longterm, sustained rise in the price of natural gas could have on its economy. A second objective of this study is to calculate the positive economic benefits that would occur if future natural gas prices were lower than had been expected previously due to the adoption of policies favorably affecting supply and demand conditions and infrastructure investments. Study Background and Assumptions Natural gas prices in California and the United States have been steadily rising in recent years. In 1999, the average annual wholesale natural gas price at the Southern California border at Topock was $2.32/million British Thermal Units ($/MMBtu), slightly greater than the Henry Hub or U.S. reference price of $2.27/MMBtu. The annual Topock Border price rose during the electricity crisis of 2001 and declined in 2002 but has been rising steadily since, with Global Insight forecasting an annual Topock border price in 2005 of $6.79/MMBtu in August 2005 (i.e., before the Gulf hurricanes Katrina and Rita). Yet due to the shut-ins of natural gas production a result of the hurricanes and continued high demand for gas for electric power generation, natural gas prices in California have risen throughout 2005 but remain lower than those in other parts of the United States. According to the Energy Information Administration (EIA), the California composite wholesale natural gas spot price (an average of the Malin, PG&E citygate, and Topock prices) on November 15, 2005 was $7.80/MMBtu, up 24% and $1.90/MMBtu from year-ago prices, and 71% greater than the price of two years ago ($4.56/MMBtu). The Topock price had risen to $7.75/MMBtu by November 15 as compared with the Henry Hub price of $9.21/MMBtu that same day. These price increases have cost California consumers and industries millions of dollars of increased costs to heat their homes, generate electricity and to provide energy for manufacturing processes. Final users of natural gas in California spent about $19.1 billion (nominal $) to buy natural gas in 2005, including $5.8 billion by households and $4.8 billion by utilities to generate electricity. Therefore, even a 5% reduction in natural gas prices would generate about $1 billion in short-term savings to consumers and industries. California is part of the North American natural gas market through its connections to the intercontinental pipeline network. Since demand, supply, and infrastructure factors in North America determine natural gas prices, California often has little direct control over market prices. Because natural gas received from U.S. Southwest production basins (i.e., San Juan and Permian Basins) constitutes the marginal supply of natural gas in California, the Southern California border price at the Topock hub is a major indicator of natural gas prices in California. The study period is the timeframe. Global Insight and the Natural Gas Study Advisory Committee which is comprised of the California Energy Commission, Pacific Gas & Electric Company, Southern California Edison Company, Southern California Gas Company and San Diego Gas & Electric Company defined three wholesale natural gas price scenarios in California using the delivered price of natural gas at Topock (i.e., called the Topock border price 1

4 in the remainder of the study) as the reference price of gas. The Advisory Committee and Global Insight also agreed that the three price scenarios should vary based on the differences in the Topock border and Henry Hub prices, enabling the study to consider the economic impact when California and U.S. wholesale natural gas prices were the same, when the California price was greater, and when the California price was less. The differences between the Topock border and Henry Hub prices were defined as the basis differences for each price scenario. The Advisory Committee decided to use the same oil price for all three scenarios (approximately $60/barrel by 2016) in nominal dollars for the WTI/NYMEX contract specification so that the economic impacts would be attributable only to the differences in the natural gas prices. The following gas price scenarios and the basis differences were used in estimating the economic impacts of alternative natural gas wholesale price levels on the California economy. The gas price scenarios are not forecasts of natural gas prices, but are intended as study parameters in order to assess the economic impacts of natural gas prices on the California economy. Low-price scenario: The average annual Topock border price will reach $5.00/MMBtu (in constant 2005 dollars) by The Henry Hub price will be $0.75/MMBtu higher than the Topock border price in 2007, a basis difference that will remain constant through Middle-price scenario: The average annual Topock border price will reach $7.50/MMBtu (in constant 2005 dollars) by No basis difference exists between the Topock border and Henry Hub prices in the Middle-price scenario. High-price scenario: The average annual Topock border price will reach $10.00/MMBtu (in constant 2005 dollars) by The Henry Hub price will be $0.25/MMBtu less than the Topock price by 2007 and then decline to $0.75/MMBtu lower than the Topock border price by 2009, a basis difference that will remain constant through The three price scenarios were selected to define the probable maximum range of future gas prices in California on an annual average basis. Global Insight and the Advisory Committee believed that future wholesale natural gas prices would most likely neither exceed the levels in the High-price scenario, nor fail to meet the levels defined in the Low-price scenario. The Middle-price scenario was not defined as a baseline or most likely forecast. Supply and Demand The California Energy Commission (CEC) forecast in November 2005 that total demand for natural gas in California will grow at an annual rate of 0.55% in the next decade. It will grow at faster annual rates of 1.76% in the Commercial sector and 1.33% in the Residential sector. By comparison, annual growth in demand by the power generation will approach but not meet the total statewide rate at 0.54%, while industrial demand will decline 0.77% annually. The supply and demand balance for natural gas in California depends on two factors: 1) sectors with the highest growth rates that drive demand and 2) the provision of new supplies. Demand growth will depend heavily on the use of natural gas in power generation, as 49% of the electricity generated in California in 2001 was produced by burning natural gas, and 41% of the electricity consumed in the state was generated by burning natural gas. The provision of new, lower-cost supply sources will depend in part upon policy options implemented in California. Adding pipeline capacity may not be enough to guarantee adequate supply, as several of the North 2

5 American basins upon which California depends for much of its natural gas face long-term production declines. Methodology Global Insight used three proprietary econometric models to estimate the economic impacts presented in this study. Three forecasting steps were required: 1) prepare three alternative U.S. energy forecasts with our Energy model using the Henry Hub and oil prices described above as inputs, 2) prepare three alternative U.S. macroeconomic forecasts using the three energy forecasts as inputs, and 3) prepare three forecasts for California using our enhanced econometric model for the State. Key outputs from the U.S. energy and macroeconomic forecasts were wholesale and retail energy prices for major types of energy, especially natural gas, along with changes in such indicators as gross domestic product (GDP), employment, industrial production, inflation, and personal income. The impacts in California were derived using Global Insight s enhanced econometric model for the state, which was revised to directly consider the U.S. and California natural gas price levels in each of the three price scenarios. We combined time series data on regional and U.S. energy prices and the energy use shares (in British thermal units or Btus) by sector to derive a weighted price of energy by sector in California. We then econometrically estimated the historic relationship between the economic growth of a sector and its weighted average price of energy in order to capture the way in which changing energy prices had historically affected levels of employment and output in the sector. Global Insight also conducted a number of interviews with representatives of companies operating in California that are major users of natural gas in order to determine how they have responded to changes in natural gas prices in the past and how they would likely do so in the future. Economic Effects of Changes in Natural Gas Prices Higher natural gas prices can affect a state economy negatively in three ways: 1) direct effects the amount end users must spend for natural gas rises; 2) indirect effects end users buy fewer inputs from suppliers because their sales decline due to price increases caused by higher natural gas prices; and 3) induced effects employment declines at the directly and indirectly affected firms lead to drops in local expenditures of disposable income. In the short run, a business has two options to offset the higher natural gas costs: pass along some or all of them to its customers; and reduce other non-gas costs by lowering employment or buying less from suppliers. In the short run, households are also forced to spend more for natural gas and less for other goods and services. Supplying firms then experience reduced sales, leading them to also reduce purchases from their suppliers, and so forth. If a manufacturing firm that spends more to buy natural gas lays off workers or reduces wages, spending at local retail stores will decline, causing them to reduce employment as sales fall, and so on. Economic Impacts in California Direct Impacts In our study, the direct economic effect is defined as the amount of money spent by final users to purchase natural gas under each of the three scenarios. In the High-price scenario, the amount of money spent will be higher than that spent in the Middle-price and Low-price scenarios or than that experienced as a cost increase by households and businesses as 3

6 compared with expenditures in the other two scenarios. By contrast, the amount of money spent to purchase natural gas in the Low-price scenario will be substantially lower than that spent in the other two scenarios, resulting in a cost savings to households and businesses. We estimated the following expenditures for natural gas under the three scenarios in Low-price scenario: Natural gas purchases of $18.8 billion (nominal $) in 2010 and $20.3 billion (nominal $) in Middle-price scenario: Natural gas purchases of $22.4 billion (nominal $) in 2010, $3.6 billion more than in the Low-price scenario; and $26.9 billion (nominal $) in 2016, $6.7 billion higher than in the Low-price scenario. High-price scenario: Natural gas purchases of $26 billion (nominal $) in 2010, $7.2 billion above the Low-price scenario; and $33.2 billion (nominal $) in 2016, $13 billion higher than in the Low-price scenario. These direct effects will spread through the California economy via the indirect and induced effects described above. The Global Insight forecast models produce estimates of the total net change in economic activity, so the direct, indirect, and induced effects are included in the results presented herein. It is important to note that the economic impacts presented in this study are cumulative and will occur with time as firms and households make continuous adjustments to natural gas price levels. The economic impacts in 2016 will be the result of adjustments made in the timeframe by the affected firms and households. Because our models use time series data, they are able to capture these adjustments over time. The economic impacts produced by our models include direct, indirect, and induced effects. Impact Summary The economic impacts in California for the three price scenarios are presented in Table ES-1, followed by impacts for the major end-user groups. The table shows that by 2016 employment in the Middle-price and High-price scenarios would total 97,700 and 163,300 fewer jobs than in the Low-price scenario. Similarly, real gross state product (GSP, in constant year 2000 dollars) in 2016 under the High-price scenario will be $30.4 billion less than that in the Low-price scenario. Global Insight finds that the best measure of the economic impacts of the three scenarios is real GSP (sometimes referred to as real value-added output) in constant year 2000 dollars as this measure is defined on the same basis as GDP and is expressed in the same base year. The differences, in numerical and percentage terms, between the levels of economic activity in the Low-, Middle- and High- scenarios from the Low-price scenario, in both numerical and percentage terms, increase during timeframe, showing that the impacts are cumulative. These widening differences are the result of adjustments made by the affected households and companies during the study period to adapt to rising natural gas prices, such as investments in more energy-efficient equipment and fuel switching capability, changes in production processes, reduced labor costs from eliminated shifts and layoffs, plant closings, and operations being shifted out of state. The adverse impacts of the Middle- and High-price scenarios by 2016 are higher for both real GSP and real personal income in part because sizable labor cost savings have already been realized in the natural gas-intensive sectors in California during the past few years; therefore, firms must find alternative ways to cope with rising natural prices other than cutting jobs. The impacts of the Middle- and High-price scenarios on real wage disbursements are larger in percentage terms than are the impacts on personal income in part because the higher gas prices will increase the rate of inflation, 4

7 Table ES-1: Summary of Economic Impacts in California Total Employment (thousands of jobs) Low Middle High 2010 Level 15, , ,716.1 Numerical Diff. from Low % Diff. from Low -0.4% -0.7% 2016 Level 17, , ,920.4 Numerical Diff. from Low % Diff. from Low -0.6% -1.0% Real Wage and Salary Disbursements (Billions of 2000$) Low Middle High 2010 Level $784.7 $780.0 $775.4 Numerical Diff. from Low -$4.7 -$9.2 % Diff. from Low -0.6% -1.2% 2016 Level $922.0 $914.2 $906.6 Numerical Diff. from Low -$7.8 -$15.3 % Diff. from Low -0.8% -1.7% Real Personal Income (Billions of 2000$) Low Middle High 2010 Level $1,540.5 $1,532.7 $1,526.4 Numerical Diff. from Low -$7.9 -$14.2 % Diff. from Low -0.5% -0.9% 2016 Level $1,844.8 $1,834.2 $1,824.0 Numerical Diff. from Low -$10.6 -$20.7 % Diff. from Low -0.6% -1.1% Real Gross State Product (Billions of 2000$) Low Middle High 2010 Level $1,810.3 $1,801.6 $1,794.5 Numerical Diff. from Low -$8.7 -$15.8 % Diff. from Low -0.5% -0.9% 2016 Level $2,327.6 $2,310.7 $2,297.2 Numerical Diff. from Low -$17.0 -$30.4 % Diff. from Low -0.7% -1.3% but more because the cumulative, adverse effects of higher gas prices lowers employment and wage and salary income. The economic benefits to the California economy of lower natural gas prices would be greater than the economic costs of higher prices primarily because the lower prices directly reduce production costs in the non-manufacturing sectors, including the costs of electricity generation, both of which can be passed along directly to customers and consumers. Lower costs would make key California industries more competitive than those in other states, resulting in lower inflation and leading to subsequent increases in real output and incomes. 5

8 Impacts by End User Households Households will be affected directly by rising natural gas prices under the High-price scenario as they spend more to purchase natural gas used for heating and cooking, and indirectly via higher electricity prices. Table ES-2 presents the changes in spending per household for natural gas and electricity under the three scenarios in nominal dollars, as households make expenditure decisions based on nominal values. Table ES-2: Changes in Household Spending for Natural Gas and Electricity Natural Gas Spending per Household (Nominal $) Low Middle High 2010 Level $440 $510 $578 Numerical Diff. from Low $70 $138 % Diff. from Low 15.9% 31.4% 2016 Level $440 $558 $673 Numerical Diff. from Low $118 $233 % Diff. from Low 26.8% 53.0% Electricity Spending per Household (Nominal $) Low Middle High 2010 Level $937 $970 $1,002 Numerical Diff. from Low $33 $65 % Diff. from Low 3.5% 6.9% 2016 Level $1,097 $1,152 $1,203 Numerical Diff. from Low $55 $106 % Diff. from Low 5.0% 9.7% Average Household Personal Income (Nominal $) Low Middle High 2010 Level $144,024 $143,623 $143,436 Numerical Diff. from Low -$401 -$588 % Diff. from Low -0.3% -0.4% 2016 Level $187,677 $187,208 $185,954 Numerical Diff. from Low -$469 -$1,723 % Diff. from Low -0.2% -0.9% In the High-price scenario, annual spending per household for natural gas would be $233 higher than it would in the Low-price scenario by 2016, due primarily to the higher Topock border price, which in turn will increase natural gas retail prices. Annual per household spending for electricity would be $106 higher, for a combined increase of $339. Consumer Expenditure Survey data for 2003 showed that about 2.2% of annual consumer expenditures (or 1.3% of average household personal income) by California households were allotted for natural gas and electricity. A major concern is the effect of the High-price scenario on low-income households (those with annual consumption expenditures of less than $25,000); in 2003 they spent more than 3.0% on natural gas and electricity and thus would be more affected by the declines in disposable household personal income. Under the Middle- and High-price scenarios per-household real disposable income will be $523 and $1,088 less than the Low-price level of $110,076 by These real income decreases are produced by the cumulative, overall declines in economic activity that will have occurred by 6

9 2016 as shown in Table ES-1 and when combined with inflationary effects of higher natural gas and electricity prices. The combination of the decreases in real household disposable income in the Middle-price and High-price scenarios and higher spending for natural gas and electricity will adversely affect California households as they must alter their spending patterns by using more of their incomes to purchase natural gas and electricity, leaving less available to buy other goods and services. Manufacturing Sector By 2016, manufacturing employment in the Middle- and High-price scenarios would account for 15,200 and 32,500 fewer jobs than in the Low-price scenario. Employment has been declining steadily in the California manufacturing sector for years; about 321,000 jobs were eliminated in We forecast that manufacturing employment will continue to decline, but at a slower rate because of the large number of jobs that have already been eliminated; by 2016, manufacturing employment would total about 22,000 fewer jobs than it does at present, falling from its current 10.4% share of total employment to 9.0%. The Middle- and High-price scenarios will accelerate a trend that is already underway, with the latter more than doubling the forecast job loss. The percentage impacts in real manufacturing GSP would be higher than those for employment; under the High-price scenario it would be 3.3% lower than in Low-price scenario by The lower impacts on employment than on real GSP indicate, as confirmed during interviews, that the natural-gas intensive sectors in the state have been responding to higher natural gas prices by lowering production costs and eliminating jobs. In the future, with most of the job cuts already taken, responses to sustained higher natural gas prices would have to come primarily through efforts to reduce the costs of production or increase productivity and energy use efficiency. Our results suggest that the price effects will predominate in the manufacturing sector, especially as energy expenditures fall, since the cost savings can be immediately captured and passed along in the form of lower prices or higher profits, or invested in other more productive ways. Private, Services-Providing Sectors The impacts on the private, services-providing (PSP) sectors in both absolute and percentage terms will be similar to the overall impacts on the total economy presented in Table ES-1, especially for the High-price scenario. The PSP sectors currently account for about 70% of total real GSP in California, so on an aggregate basis, most of the impacts of higher natural gas prices will occur in these sectors. The primary economic effects of these prices will increase the costs of occupying office space, including electricity. We estimate that in 2016, real GSP in the PSP sectors in the High-price scenario will be $16.3 billion less than in the Low-price scenario, or about 54% of the total difference in real GSP. This lower impact share is unsurprising, as PSP businesses will be less affected by higher natural gas prices than manufacturing companies will be. Government Sector The economic impacts of gas prices on the government sector will be relatively small, again because the primary effect will be seen in the costs of occupying office space and using electricity. Some exceptions exist, however; notably, government agencies engaged in the generation, purchase, and sale of electric power, and Metropolitan Transportation agencies that purchase large amounts of electric power to operate subways. Government agencies most susceptible to changes in natural gas wholesale prices include those that enter into long-term 7

10 contracts to purchase either large amounts of natural gas for generating electricity or large amounts of wholesale electric power, with the electric power later sold at retail. We forecast that real GSP in the government sector (Federal, State, and Local) in 2016 under the High-price Scenario will be about $0.8 billion less than in the Low-price Scenario. Electricity Generation One of the primary direct effects of alternative natural gas prices will be on the cost of generating electricity because California generates a much higher share of its electricity using natural gas than does the rest of the country. According to the EIA, in 2001, slightly more than 49% of electric power generated in California was produced by burning natural gas; only 1.0% was made by burning coal, as compared with U.S. shares of 16.4% and 51.2%, respectively. Higher natural gas prices will directly raise the price of generating electricity, requiring businesses and households to pay more to obtain it and providing them an incentive to use less of it. We estimate that by 2016, the cost of natural gas purchased by California electric utilities and used in generating electric power in the Middle- and High-price scenarios will be $6.5 billion and $8.3 billion greater than in the Low-price Scenario, producing higher retail electricity prices and leading to the aforementioned negative impact on households. Impacts in Northern and Southern California Global Insight estimated the impacts of high natural gas prices on Northern and Southern California regions; the latter was comprised of nine counties (Imperial, Los Angeles, San Luis Obispo, Santa Barbara, Orange, San Diego, Ventura, Riverside, and San Bernardino) and portions of four others (Kern, Tulare, Kings, and Fresno) based on utility service areas. The Northern California region covered the remainder of the state and generally aligns with the natural gas service area of the Pacific Gas and Electric Company (PG&E). Economic impacts for most indicators, such as real GSP and employment, will be slightly larger in Southern California than in Northern California because the former has higher concentrations of nine manufacturing sectors that are heavy users of natural gas. As a result, the Southern California economy is more sensitive to natural gas prices because of its economic structure; it would benefit more from lower natural gas prices and would be affected more adversely by higher prices. Findings Level of State-wide Economic Activity Falls with Higher Prices Natural gas prices under the Middle- and High-price scenarios would result in lower levels of economic activity in California by 2016 when compared with the levels in the Low-price scenario. Using outputs from our Energy and U.S. Macroeconomic models, the simulations performed with our enhanced California econometric model showed that sizable declines in the level of economic activity would be observed by 2016 in the High-price scenario, notably a $30.4 billion drop in real GSP ($42.2 billion in nominal terms) and the accompanying loss of 163,300 jobs. In percentage terms, the differences in levels of economic activity (as measured by total employment, total real GSP, and total real personal income) between the Low- and High-price Scenarios would range between 1.0 and 1.3% due to the large size of the California economy (currently 13.3% of US GDP). The percentage impacts would be higher in the manufacturing sector, where we forecast that by 2016, employment and real GSP would be 2.1% and 3.3% lower, respectively, in the High-price scenario than in the Low-price scenario. 8

11 Households Affected by Higher Energy Expenditures and Lower Incomes California households would be adversely affected by sustained higher natural gas prices in several ways: 1) by the declines in the levels of economic activity, especially the loss of employment and the corresponding drop in wage and salary income; 2) by higher spending for natural gas and electricity, which reduces money available to purchase other goods and services; and 3) by higher inflation resulting from the increased nominal prices for energy, which in turn lowers real disposable personal income. We estimate that combined expenditures for natural gas and electricity would total $1,876 per household in nominal terms in the High-price scenario by 2016: $1,203 for electricity and $673 for natural gas. Increased spending by households on natural gas and electricity means they would have to limit their purchases of other goods and services. The percentage decline in the real wage and salary earnings in the High-price scenario as compared with that in Low-price scenario would be 1.7% by 2016, or far greater than the percentage drop in total real personal income, due to the combined effects of lower employment and higher price levels. Marginal Impacts Decline as Natural Gas Prices Rise Marginal economic impacts would decline as natural gas prices rise and increase as gas prices fall. The marginal economic impact is the change in the level of economic activity (e.g., employment, income, GSP) produced by a direct economic effect, such as an increase or decrease in the price of natural gas. The existing price level when a direct effect occurs largely determines the size of the marginal economic impact: if the natural gas price level is already high a further price rise of 1% would have a smaller, negative, marginal economic impact than a 1% increase from a low price level. Put in the context of this study, the number of jobs lost for each 1% increase in price between the Middle and High-price scenarios would be less than the number lost for each 1% increase in price between the Low- and Middle-price scenarios. The reverse would also occur: the increase in employment for each 1% decrease in price between the High- and Middle-price scenarios would be less than the increase in employment for each 1% decrease between the Middle- and Low-price scenarios. The declining marginal economic effects of rising natural gas prices were identified by estimating implicit employment elasticities (defined as the percent decline in employment for each 1% increase in natural gas prices between two scenarios). Amid rising prices, the decline in marginal impacts means that up to a certain price level, most of the adjustments that can be made to reduce energy costs will have been implemented and the savings realized; beyond this price, little more can be done to lower energy use without also lowering production levels (i.e., no more workers can be laid off, energy use cannot be reduced any further, and all energy-efficient equipment has been installed). The energy cost savings that can be realized will become smaller, and the marginal costs required to obtain it will become increasingly higher. With falling prices, such as from the High- to the Low-price scenario, the reduction in natural gas and electricity expenditures are immediately realized, enabling businesses to either lower prices, increase profits, or invest the freed-up funds in more productive ways. At the same time, households have more money to spend on other goods and services. Impacts of Higher Natural Gas Prices Increase in Time Our study showed that the adverse economic impacts of the High- and Middle-price scenarios will increase with time in both absolute (e.g., using constant year 2000 dollars for GSP and income) and percentage terms when compared with those of the Low-price scenario. The primary reason is that the price levels for the three scenarios, in real terms, widen over time. 9

12 Businesses and households directly affected by sustained higher prices will continually adjust to sustained higher natural gas prices in order to reduce the amount they spend for natural gas and electricity, so that in time the differences in the levels of economic activity among the three scenarios will widen. This process of continual adjustment also means that by 2016 the differences in the levels of economic among the three price scenarios will be the cumulative result of all the adjustments that were made in the preceding 10 years. The tables presented in Section 5.0 and in Appendix A consistently show larger absolute and percent impacts in 2016 than they show in California is More Sensitive to High Natural Gas Prices than the U.S. Economy Our study showed that the California economy is somewhat more sensitive to higher natural gas prices than is the U.S. economy. We corrected for differences in the levels of wholesale natural gas prices in the United States and in California under the three scenarios by estimating implicit elasticities for such key indicators such as real GDP/GSP, employment, and personal income. The elasticities consistently showed that a 1% increase in the wholesale price of natural gas in the United States would generally have a smaller, adverse impact in percentage terms than a 1% increase in wholesale natural gas prices in California. The California economy is more sensitive to natural gas prices for a couple of reasons: 1) its major natural gas using industries obtain higher shares of their energy inputs (on a Btu basis) from natural gas than do the same sectors at the U.S. level; and 2) almost half of the electric power generated in California comes from burning natural gas, resulting in higher retail electricity prices for the major end-user groups, such that the negative effects of the higher energy expenditures will spread throughout the economy. Higher residential electricity prices are significant because California households currently spend about $1.91 for electricity for every $1.00 spent for natural gas. Structure of California s Economy would be Unaffected The composition or structure of a state economy is shown by the types of goods and services it produces, and is usually measured by the relative shares of economic activity, such as employment and output, by economic sector (i.e., what percent of state-wide employment and real GSP is in the manufacturing sector). A number of domestic, regional, and global economic forces have and will continue to change the structures of both the U.S. and California economies. Many of these factors are well-known, including: The continuing decline of the manufacturing sector in California and throughout the United States Increased global competition, such as the rises of China and India Off-shoring Current and future trade agreements, such as NAFTA The competitive advantages and disadvantages of the California economy, such as high labor and housing costs in its largest MSAs Regional migration trends of households and businesses in the western United States California business taxes and environmental and energy policies Although these factors will continue to effect economic growth in California, we conclude that, when considered in the context of all these powerful factors, higher sustained natural gas prices would have a minor impact on the structure of the California economy. The structure of the state economy will continue to evolve due to the aforementioned factors; the primary structural effect of higher natural gas prices would be to accelerate the decline of the manufacturing sector, but this downward trend will continue independent of the future price of wholesale natural gas. 10

13 Conclusion Our study shows that sustained, higher natural gas prices in the High- and Middle-price scenarios would have negative effects on California households and commercial and industrial end-users, resulting in cumulative reductions in the forecasted levels of employment, real GSP, and income by Higher natural gas prices would also accelerate the continuing loss of manufacturing jobs, produce higher utility bills for natural gas and electricity customers, and result in lower levels of household income. By contrast, lower natural gas prices in the Low-price scenario would benefit directly residential, industrial, and commercial customers alike, as they would spend less to purchase natural gas and electricity, resulting in higher levels of employment, GSP, and real incomes by

14 INTRODUCTION Natural gas prices in California are currently rising, although they remain lower than those in most other parts of the United States. According to the Energy Information Administration (EIA), the California composite wholesale natural gas spot price (an average of the Malin, PG&E citygate, and Southern California border price at Topock) on November 15, 2005 was $7.80/million British Thermal Units ($/MMBtu), up 24% and $1.90/MMBtu from year-ago prices, and 71% greater than the price seen two years ago ($4.56/MMBtu). The Topock border price had climbed to as much as $11.67/MMBtu in late October By contrast, the Henry Hub natural gas price on November 15 was $9.21/MMBtu. The final users of natural gas in California will spend about $16.6 billion to buy natural gas in 2005, so even a 5% reduction in its price would generate a considerable short-term savings for companies, utilities, and households. California is part of the North American natural gas market through its connections to the intercontinental pipeline network. Demand, supply, and infrastructure factors in North America determine natural gas prices, so California often has little direct control over market prices. For example, in February 2003, a time when California demand was moderate, California natural gas wholesale prices spiked due to extreme weather conditions in the Northeast. Such is not the situation today, however: the increasing national hunt for a limited supply of natural gas is driving prices higher, and therefore, California must focus on those actions that can help California consumers and that are within its control. Natural gas sent from southwest production areas constitutes its marginal supply, and thus the Topock Border price is an indicator of natural gas prices in California. Study Objectives The objective of this study was to assess the impacts of changes in natural gas prices on the California economy. Because California is more dependent on natural gas as a source of energy than are most other states, the adverse economic effect that a continuing, long-term rise in the price of natural gas could have on the statewide economy is a major concern, especially if California natural gas prices begin to exceed U.S. prices. It was also important to estimate the economic benefits that could occur if future natural gas prices were lower than had been expected through favorable supply and demand conditions and infrastructure investments. Requirements of the Study Sponsors The sponsors of the study (listed in the next section) had the following requirements that determined the scope of the work and the methodologies applied. Use publicly available economic and gas price data to conduct the study. Make study methodology, data sources, and results available to the public. Measure macroeconomic impacts of various gas price scenarios on income, employment, key industrial users of gas, in Northern California, Southern California, and the state economy (GSP) as a whole. Representatives of its sponsoring organizations were asked to serve on an advisory committee for the study; its purpose was to determine the scope of the study, review and approve study approaches, provide technical guidance and input since all its members were knowledgeable about the California natural gas sector, and review and provide comments on interim deliverables and the draft and final reports. Members of the Advisory Committee included: 12

15 Les Buchner, Pacific Gas and Electric Company Robert Cowden, Pacific Gas and Electric Company Herb Emmrich, Southern California Gas Company and San Diego Gas & Electric Company Jairam Gopal, California Energy Commission Karen Griffin, California Energy Commission Richard Hendrix, Pacific Gas and Electric Company Dave Maul, California Energy Commission Luis Pando, Southern California Edison Company Scott Wilder, Southern California Gas Company and San Diego Gas & Electric Company 13

16 BACKGROUND Natural Gas Supply and Demand in California California is literally at the end of the natural gas pipeline, and as such has experienced periods of both oversupply and undersupply relative to its demand growth. Also, a growing shortfall of the natural gas commodity is itself creating repercussions in gas prices. The loss of production from the 2005 hurricane season will amount to more than 700 billion cubic feet (bcf), constraining potential users of natural gas and creating a very high price environment. Although such disruptions are hopefully unique and temporary in their effects, they emphasize the stresses of operating in a natural gas market and highlight concerns as to whether sufficient supplies of natural gas and adequate infrastructure will be available in the future. Simply adding pipeline capacity is not enough to guarantee security of supply, as several of the North American basins upon which California relies are declining in the long term, and natural gas may not be available in the quantity the state would require. Also, although the current infrastructure appears adequate, the availability of these supplies may be insufficient to meet demand at all times. Demand The California Energy Commission (CEC) forecast in November 2005 that total demand for natural gas in California will grow at an annual rate of 0.55% in , and at higher annual rates of 1.76% in the commercial sector and 1.33% in the residential sector. By contrast, annual demand growth for power generation will be 0.54%, while industrial demand will decline 0.77% annually. These growth rates are similar to those calculated by Global Insight in its forecast, which also shows a small decline in industrial natural gas demand. Figures 1 and 2 present comparisons of natural gas demand growth rates prepared by Global Insight, the CEC, and the EIA. Global Insight has undertaken a study focusing upon the price sensitivity of industrial natural gas demand during this period. Industrial demand includes all energy-intensive manufacturing sectors (such as petroleum refining) that together account for the majority of demand. In the middle-price scenario, California natural gas demand will grow 0.6% in the core sectors and 0.8% in the power-generation sector, but will decline 0.1% in the industrial sector. The Energy Information Agency (EIA) Pacific region forecast calls for even higher growth prospects for California than does Global Insight. The supply and demand balance for natural gas in California is dependent upon two factors: those causing the highest growth and those relating to new supplies. Demand growth results from the use of gas in power generation. The provision of new, lower-cost sources of supply reflects various policy options within California. If a miscalculation is made, it is likely that the industrial sector will suffer the greatest consequences. When the balance of supply and demand is tight, such as it was in , it is often the industrial sector that carries the burden of reducing demand to meet supply-side constraints. 14

17 7,500 Figure 1: Alternative Natural Gas Demand Forecasts for California 7,250 7,000 (MMCF/DAY) 6,750 6,500 6,250 6,000 5,750 5, California Energy Commission EIA History Global Insight Figure 2: Comparison of Global Insight and EIA Forecasts of Pacific Natural Gas Demand (California, Oregon, Washington) 9,500 9,000 8,500 (MMCF/DAY) 8,000 7,500 7,000 EIA Forecast , Excludes Lease, Plant and Pipeline Fuel 6, Energy Information Administration Global Insight 15

18 Electric power generation in California is highly dependent upon the use of natural gas as a fuel, and as a result, consumer electricity prices are very sensitive to the cost of natural gas. According to the EIA, natural gas was used to generate about 49% of the electricity produced within the State in 2001, and 41% of all electricity used there. About 20% of the energy used in California comes from out-of-state sources, with virtually all imported electricity coming from sources that do not burn natural gas, such as coal-fired power plants and hydroelectric facilities. Industrial customers in California use large amounts of natural gas to produce steam and cogenerate electricity which is then sold to the state s electric utilities. The EIA divides this type of industrial natural gas demand into two parts: 1) the portion used to produce steam, which is assigned to industrial natural gas demand; and the 2) the portion used to generate electricity, which is assigned to electric generation (EG) natural gas demand. Global Insight uses these same definitions because we use EIA data in our energy model. By contrast, the CEC defines the entire amount of the natural gas used by industrial customers to produce steam and cogenerate electricity as electric generation (EG) demand. The result of these differences is that although Global Insight and the CEC have similar totals for industrial and electric generation natural gas demand, our share for industrial demand is higher than that of the CEC while our share for electric generation demand is lower. A significant share of the electric power produced in California comes from co-generators, so that higher industrial natural gas prices would increase the cost of electricity produced by co-generators, which would increase electricity rates for consumers, depending on the extent to which higher natural gas costs are passed along to the consumers. A discussion of the differences in the definition of industrial demand and how they would affect the relationship between natural gas and electricity prices is presented in the Merchant Power section of Appendix B, which contains the results of our interviews with major natural gas users. Supply The existing California supply portfolio has evolved over time from strictly local supplies to now include the Southwestern United States and such areas as the Permian, Anadarko, and San Juan basins, Alberta (after the completion of the PGT pipeline), and most recently, the Kern River pipeline from the Rocky Mountains. By 2008, California will use a small amount of LNG in its natural gas mix to further enhance its regional supply diversity. None of these supplies would be possible without the infrastructure such as pipelines to bring the gas commodity from as far as a thousand miles away. Pipeline expansions and new construction have kept up with California demand growth but in starts and fits. During late 2000, many difficulties arose in the western natural gas markets such that in 2001, an expedited construction of new pipeline capacity helped align the natural gas markets back towards normalcy. This program featured a doubling of the capacity of the Kern River Pipeline that connects Wyoming gas supplies to the California market. Other pipelines will eventually bring new supplies into the California market. The Bajanorte pipeline, which currently runs from Ehrenberg on the Southern California border with Arizona to Baja, Mexico, will be reversed in 2008 to access the Energia Costa Azul LNG terminal in Mexico. Construction has begun on this terminal, and it is expected to be completed by mid Also, contracts for LNG have been signed by project sponsors for Australian, Russian, and Indonesian supplies. Some of the LNG from the Baja region would likely be shipped into California. Other supply projects have also been proposed that could eventually enhance the regional diversity of California supply, and could influence the trend in natural gas prices. This study 16

19 does not specifically refer to any such new infrastructure, but rather examines the impact of changing natural gas prices on the industrial sector. Recent Price Trends Figure 3 shows the recent trends in the average annual price levels for the Topock Border and Henry Hub price levels. The annual prices for 2005 are based on year-to-date price information through late November The figure shows the steady rise in the average annual wholesale natural prices in California and the United States. Whether the Topock Border price will exceed the Henry Hub price for a sustained period remains a future concern for California; as Figure 3 shows the Topock border price was higher than the Henry Hub price in Figure 3 also shows that in recent months, the basis differential between the Topock border and Henry Hub price has increased. Figure 3: Recent Trends in the Topock Border and Henry Hub Natural Gas Price Levels $10.00 $9.00 Price in $/MMBtu (Nominal $) $8.00 $7.00 $6.00 $5.00 $4.00 $3.00 $2.00 $1.00 $ Topock Price Henry Hub Price Looking at more recent price trends, natural gas prices in California are currently rising; although they remain lower than in most other parts of the United States. According to the Energy Information Administration (EIA), the California composite wholesale natural gas spot price (an average of the Malin, PG&E citygate, and Southern California border price at Topock) on November 15 was $7.80 per million British Thermal Units ($/MMBtu), up 24% and $1.90/MMBtu from a year ago, and 71% greater than the price of $4.56/MMBtu from two years ago. The Topock Border price reached as high as $11.67/MMBtu in late October By contrast, the Henry Hub natural gas price on November 15 t was $9.21/MMBtu. The final users of 17

20 natural gas in California will spend about $16.6 billion to buy natural gas in 2005, so that even a 5% reduction in its price would generate a considerable short-term savings to companies, utilities, and households. California is part of the North American natural gas market through its connections to the intercontinental pipeline network. Demand, supply, and infrastructure factors in North America determine natural gas prices, so California often has little direct control over market prices. Because natural gas sent from southwest production areas constitutes the marginal supply, the Topock Border price is a major indicator of natural gas prices in California. 18

21 STUDY ASSUMPTIONS This study was undertaken before, during, and after Hurricanes Denis, Katrina, Rita, and Wilma ravaged the U.S. Gulf Coast from August to October These pricing scenarios were chosen before the hurricanes hit in order to abstract from a specific forecast a wide range of possible price futures available to California based upon prudent management of its natural gas supply and demand. An average, nominal oil price of $60/barrel and an average real natural gas price of $7.50/MMBtu (2005$) were chosen for the Middle-price scenario. A nominal price of $60/barrel in this period would be about $53/barrel in real terms; noting that oil futures are traded in nominal values. To assure such a gas price for California would require adding infrastructure to meet growing demand. Natural Gas Price Scenarios In conjunction with the California Advisory Group, the Global Insight team prepared three plausible price scenarios for wholesale natural gas during a 10-year forecast horizon from 2006 through The three price scenarios referred to in this study as the Middle-, High-, and Low-price scenarios were prepared initially for the natural gas border price for the state of California, and then used to derive all other natural gas wholesale and retail prices for both the United States and the state of California. The next step was to use these natural gas price scenarios in conjunction with oil and other energy price forecast to analyze the impact on the economies of California and the United States. The natural gas price scenarios were developed in cooperation with the Advisory Committee to reflect a consensus on levels of prices, regional differences in price, and rates of price change over time. The purpose of the price scenarios was not to forecast prices, but rather to reach a consensus on the price scenarios to be used in examining economic impacts, such as variations in the levels, rates of change, and basis differences that compare California with other regions in the United States. The Advisory Committee decided to use the same oil price scenario for all three cases so that the impact of each would be directly attributable to the natural gas price scenarios. The crude oil price was derived from the mid-august 2005 futures strip through 2011 extended to 2016 while keeping the 2011 value constant. The crude oil price is approximately $60 nominal for the WTI/NYMEX contract specification. Also, a difference in the basis between Henry Hub (which we used as the measure of the U.S. wholesale price for natural gas) and the Southern California Border price at Topock was assumed to be either $0.00 plus or minus $0.75/MMBtu to assist in investigating the relative importance of price level versus basis for California economic impacts. A major concern of the Advisory Group was that the prices used in this study should not be easily identifiable as originating from a specific forecast; rather, that they ought to be based on publicly available information as part of a scenario. Finally, the ratio of gas to oil prices should be plausible. In order to accommodate these competing objectives, the decision was made to use values for the futures market where possible. The basis to Henry Hub was prepared and used to calculate the Henry Hub prices used for the Global Insight energy model based on their differences from the assumed wholesale natural gas prices in California. Thus the basis was set at +/-$0.75/MMBtu relative to a $0.00 value for the middle or baseline scenario. With the assumption, based on historical data, that California 19

22 prices will be more volatile than will U.S. prices, Henry Hub prices are lower than California prices in the High-price scenario and higher in the Low-price scenario. The California border price is thus arithmetically equivalent to the sum of the Henry Hub and basis prices. The three price scenarios are summarized below. Low-price scenario: the Topock Border price will reach $5.00/MMBtu (in constant 2005 dollars) by The Henry Hub price will be $0.75/MMBtu higher than the Topock Border price in 2007, a basis difference that will remain constant through Middle-price scenario: the Topock Border price will reach $7.50/MMBtu (in constant 2005 dollars) by No basis difference exists between the Topock Border and Henry Hub prices in the Middle-price scenario. High-price scenario: The Topock Border price will reach $10.00/MMBtu (in constant 2005 dollars) by The Henry Hub price will be $0.25/MMBtu less than the Topock Border price by 2007 and decline to $0.75/MMBtu less by 2009, a basis difference that will remain constant through California prices were expressed in real terms as the starting point for calculating Henry Hub prices. Henry Hub prices are defined as the California price as adjusted by the assumed basis. The Topock Border basis is set at $0.75 more than Henry Hub for the High-price scenario and $0.75 less than Henry Hub prices for the Low-price scenario. These values are phased-in to retain a separation in the Henry Hub prices. Figure 4 presents the basis for the three scenarios. The historical data for the California and Henry Hub prices presented in Figure 4 show that California natural gas prices are more volatile than Henry Hub prices. Figure 4: Basis Difference between Henry Hub and Topock Border Prices $1.25 Henry Hub-Topock in Nominal $/MMBtu $1.00 $0.75 $0.50 $0.25 $0.00 -$0.25 -$0.50 -$0.75 -$1.00 -$ Low Middle High 20

23 Finally, Figure 5 presents prices for the three scenarios in $/MMBtu in constant 2005 dollars. The real price comparison presented below in Figure 5 shows the assumption of continual increases in real prices in the High-price scenario and continuing decreases in the Low-price case. Henry Hub prices also demonstrate this pattern, but with less volatility than that shown for California. The real basis difference is -$0.75/MMBtu in the High-price scenario and +$0.75/MMBtu in the Low-price scenario for a total contribution of $1.50 towards the diversion in California Border prices. A basis of $0.00 is used for the middle scenario. Figure 5: Topock Border and Henry Hub Real Prices and Basis by Scenario $11.00 $10.00 $9.00 $2005/mmbtu $8.00 $7.00 $6.00 $5.00 $ Topock - High Henry Hub - High Henry Hub - Middle Topock - Middle Henry Hub - Low Topock - Low Figure 6 presents the price forecasts for the three scenarios in nominal dollars per MMBtu. The nominal Topock Border price in each of the three scenarios in 2016 would be: Low-price scenario, $6.41/MMBtu; Middle-price scenario, $9.61/MMBtu; and High-price scenario, $12.81/MMBtu. Crude Oil and Natural Gas Prices During times of stress in either the oil or gas markets, the relative prices of each expressed in equivalent terms or $/MMBtu can change greatly. During a decade-long period, however, natural gas prices will most likely be less than the energy equivalence of crude oil. This price relationship is very important in those markets that can switch between natural gas and petroleum products (such as oil/gas-fired power plants in the eastern United States) or for those items that can use either natural gas or petroleum as a raw material (such as ethylene plants in the U.S. gulf coast). No direct substitution will occur in California, so local prices do vary. 21

24 Figure 6: California Topock Border Price by Scenario in Nominal $/MMBtu $14.00 $13.00 $12.00 Price in Nominal $/MMBtu $11.00 $10.00 $9.00 $8.00 $7.00 $6.00 $5.00 $4.00 $3.00 $ Low Price Scenario Middle Price Scenario High Price Scenario Both crude oil and natural gas prices have been volatile in 2005: crude oil prices have increased $1 2/month for the past two years, and continued to do so in early August 2005 when the U.S. energy and macroeconomic simulations used in this study were performed. The mid-august futures strip boasted a value of about $60/barrel in nominal terms. Figure 7 presents the oil price assumption used in this study. A key implication for the crude oil and natural gas assumptions is the ratio of gas to oil prices in energy equivalent terms, or $/MMBtu. When gas prices exceed oil prices, it affects the economic rationale of such activities as enhanced oil recovery and petrochemical feedstock for ethylene. Figure 8 below presents the ratios of the Henry Hub to crude oil prices for the three natural gas price scenarios used in this study. Even when offering less volatility for Henry Hub prices, the Henry Hub gas price in the High-price scenario will rise to 117% of the crude oil price by California gas prices would approach 125% of crude oil prices by 2016 in the Highprice scenario. Economic Sectors Dependent on Natural Gas A major determinant of the economic impacts of alternative natural gas prices will be the number, type, and size (i.e., as measured by levels of valued added, employment, or income) of the economic sectors in California that are dependent on natural gas. We define natural gasdependent sectors as those principally in manufacturing but also including electric power generation and some transportation activities that use large amounts of natural gas, and where the cost of natural gas is a significant share of the value of the output. In 2003, Global Insight prepared the following report: Demand Destruction the Impact of Rising Natural Gas Prices. The study identified the following industries at the U.S. level that are major users of natural gas and, thus, are those most likely to be adversely affected by rising prices, with their three-digit 22

25 Figure 7: WTI History and NYMEX Strip Price Mid August 2005 $70 $65 $60 Price of Oil in $/Barrel $55 $50 $45 $40 $35 $30 $25 $ Figure 8: Ratio of Natural Gas (Henry Hub) to Crude Oil Prices (WTI NYMEX Strip as of August 9, 2005) Pricd Ratio Low Middle High 23

26 NAICs codes and the cost of energy as a share of the value of output (obtained from the 1997 benchmark input/output use table for the U.S. economy presented in parentheses): Ethylene and Propylene (325 Chemical Manufacturing, energy use is 3.9% of the value of output) Ammonia (325 Chemical Manufacturing, energy use is 3.9% of the value of output) Methanol (325 Chemical Manufacturing, energy use is 3.9% of the value of output) Other Chemicals (325 Chemical Manufacturing, energy use is 3.9% of the value of output) Cement (327 Non-Metallic Mineral Products Manufacturing, energy use is 4.3% of the value of output) Aluminum (331 Primary Metals; some products such as forgings are in 332 Fabricated Metals, energy use is 4.9% of the value of output) Steel (331 Primary Metals, some products such as forgings are in 332 Fabricated Metals, energy use is 4.95% of the value of output) Paper (332 Paper and Paper Products, energy use is 1.5% of the value of output) Glass (327 Non-Metallic Mineral Products, energy use is 4.3% of the value of output) Food (311 Food Manufacturing, energy use is 1.3% of the value of output) Petroleum Refining (324 Petroleum and Coal Products, energy use is 11.6% of the value of output) Electricity Generation (221 Electric Power Generation and Supply, energy use is 8.1% of the value of output) The study noted that power generation sector has become increasingly reliant on natural gas as a fuel, so that the cost of producing electricity has become increasingly dependent on the price of natural gas. The first four sectors listed above use natural gas primarily as a feedstock, while the other sectors primarily use natural gas as fuel to generate process steam and heat. Increases in the cost of the above goods affect the cost of other goods in which they are used as an input; for example, ethylene is heavily used in making plastic products such polyethylene, which in turn affects the cost of plastic wrap used in packaging. Similarly, ammonia is used to make fertilizer, so that an increase in the price of ammonia ultimately raises the costs of agriculture production. The extent to which the California economy contains significant shares of economic activity in the above sectors will determine, to a significant extent, how the state would be affected by rising natural gas prices. Figure 9 below presents natural gas costs as a percent share of sales or output for natural gas intensive sectors. The table clearly shows the importance of natural gas prices to the cost of production in the fertilizer, petrochemical, plastics, basic chemicals, alumina, and alkalis sectors, all of whose cost shares exceed 10%. The demand destruction study described four ways in which a natural-gas dependent industry would respond to higher natural gas prices: Interfuel substitution and conservation: initial reaction to prices including temporary changes in energy usage. These are responses based on utilization of existing equipment. Technology change and efficiency improvements: long-term adjustment to prices including permanent changes in energy usage. The efficiency improvements can include investments to modify or replace equipment. 24

27 Operational changes including shutdowns: temporary or permanent reductions in output Relocation or displacement: new investment in energy-intensive industries in overseas locations and/or importation of energy-intensive components. These four ways of responding to higher natural gas prices will be implemented differently in the short and long term, so that as noted below, the effects of sustained higher natural gas prices will be cumulative over time. As explained below, a key step in revising our California economic model for use in this study was to first identify the shares of energy used by major energy type (e.g., natural gas, petroleum, coal, and electricity) in Btu by economic sector, focusing on the natural-gas intensive sectors. Our objective was to identify the economic sectors in California that were Figure 9: Natural Gas Costs as a Percent Share of Sales Nitrogeneous Fertilizer Petrochemical-LPG Plastics - LPG Basic Chem - LPG Alumina Alkalies Brick Glass Carbon Gypsum Glass Containers Organic Chemicals Other Glass Petrochemicals Ethyl Alcohol Paperboard Iron&Steel Plastics Lime Wood Chemicals Inorganic Paperboard Newsprint Petroleum Refining Cyclic Crude Glass Products Sawmills 28.30% 21.40% 16.40% 13.50% 12.40% 9.50% 8.40% 8.30% 7.00% 5.90% 5.30% 4.30% 4.00% 3.60% 3.50% 3.50% 3.00% 2.80% 2.10% 2.00% Note: Adding LPG's, Petrochemicals share increases to 28.3%, 1.90% Plastics 21.4% and Other basic chemicals, 16.4% 1.70% Source: U.S. Department of Commerce, 2002 data 1.60% 1.00% 0.60% 0.20% 45.30% 0.0% 5.0% 10.0% 15.0% 20.0% 25.0% 30.0% 35.0% 40.0% 45.0% 50.0% Natural Gas Cost as % of Sales dependent on natural gas, which we defined as sectors 1) where the value of natural gas consumed (i.e., amount in BTU consumed times the appropriate price) comprised a significant share of the cost of production; and 2) that produced significant shares of total state gross output. We used the results of our Demand Destruction Study, along with information from the Energy Information Administration (EIA), the CEC, input/output (I/O) coefficients published by the Bureau of Economic Analysis, and a proprietary data base on gross output by sector in 25

28 California to identify the following economic sectors in California as being dependent on the price of natural gas. We identified the following natural gas-intensive economic sectors, with the NAICs code presented in parentheses. Food Manufacturing (311) Beverage and Tobacco Products (312) Other Chemicals (325 Chemical Manufacturing) Cement (327 Non-Metallic Mineral Products Manufacturing) Aluminum (331 Primary Metals; some products such as forgings are in 332 Fabricated Metals) Steel (331 Primary Metals, some products such as forgings are in 332 Fabricated Metals) Paper (332 Paper and Paper Products) Glass (327 Non-Metallic Mineral Products) Food (311 Food Manufacturing) Petroleum Refining (324 Petroleum and Coal Products) Electricity Generation (221 Electric Power Generation) Many of the industries contained in the private, services-providing sectors, such as real estate, rental and leasing and retail consume large amounts of natural gas in the aggregate for space heating and hot water use, and also use large amounts of electricity, but the costs of the natural gas and electricity are low shares of the value of output produced, so they were not considered to be natural-gas intensive. 26

29 STUDY METHODOLOGY We found one study that also measured the economic impact of the higher natural gas prices on the state of California. This study was done by Professor Philip J. Romero of University of Oregon. It did not use a detailed modeling framework to calculate the impact of natural gas prices, but instead used a simple co-efficient of price elasticity between total Gross State Product and the price of natural gas. Furthermore, the impact calculated using the simple price elasticity co-efficient without use of an economy-wide model may not account for indirect effects and induced effects of changes in the natural gas price and the total impact would therefore be underestimated. Global Insight has applied a methodology that uses three different sets of models and made sure that all the interactions in the energy markets of the United States and California have been incorporated in assessing the impact of changes in natural gas prices on the economy of California. These models employed thousands of equations and time-series data on thousands of economic concepts. The final result of this approach yields dynamic net impacts of changes in the energy prices. In addition to estimating the impact using econometric models, Global Insight also conducted a number of interviews with representatives of companies operating in California that are major users of natural gas in order to determine how they have responded to changes in natural gas prices in the past, and how they would likely do so in the future. Most of the companies were manufacturers, but we also contacted firms in the non-manufacturing sectors (such as power production) that procure electric power under long-term contracts, as well as engineering, construction, and energy efficiency equipment supplier firms. Since the objective of this study was to assess the impacts of changes in natural gas prices on the California economy, the starting point was the three wholesale natural gas price scenarios developed by the Advisory Committee and Global Insight. The next step was to use these natural gas price scenarios along with assumptions on the price of oil to first determine the potential impact of these predictions on the U.S. economy and then the impact of each of these on the economy of California. Our methodology consisted of preparing alternative economic forecasts through 2016 in the following three steps. Preparing Energy Forecasts Global Insight and the Advisory Group first identified price levels for California wholesale natural gas during for three scenarios, and then derived the resulting Henry Hub prices based on the +$0.75/MMBtu and -$0.75/MMBtu basis differences. The Global Insight Energy Service used its U.S. Energy Model to produce three alternative U.S. energy forecasts that incorporated the middle, high, and low natural gas price scenarios. For example, a separate U.S. energy forecast was first generated using the Middle-price scenario in which there was no basis difference between the California border price and the U.S. wholesale natural gas price as indicated by the Henry Hub price. Separate U.S. energy forecasts were then prepared for each of the other two scenarios. We next assumed that the price of oil would be the same under all three scenarios. With these energy prices set, we then prepared three energy forecasts for the United States and Department of Energy (DOE) regions, with each forecast including forecasts of retail natural gas prices, and electricity prices in California and the U.S. by major end user (i.e., residential, commercial, industrial, or electric power generator). This approach meant that the resulting forecasts of the retail prices of natural gas and electricity in California were a function of the initial assumptions of wholesale natural gas prices in California. 27

30 Estimating U.S. Macroeconomic Impacts The Global Insight macro model of the U.S. economy has the capabilities to receive inputs from the energy modeling system and to perform the simulation. This model is an econometrically estimated equilibrium growth model and treats U.S. economy as one entity. It is a standalone model of the U.S. economy and therefore is useful to measure a net impact on the national economy. The U.S. macroeconomic model captures the full simultaneity of the U.S. economy, spanning final demands, aggregate supply, prices, incomes, international trade, industrial detail, interest rates, and financial flows. The outputs from the three energy forecasts were used as exogenous variables in the Global Insight U.S. macroeconomic model to compute the U.S. macroeconomic impact of the Middle-price, High-price, and Low-price scenarios. The results of these three simulations showed the effects of alternative natural gas prices on such important macroeconomic variables as gross domestic product (GDP), industrial production by sector, employment by sector, consumption, investment, etc. Estimating California Economic Impacts The final step in our methodology was to prepare an estimate of the impact of the three natural gas price scenarios on the California economy, using as inputs the results from the preceding steps. Key outputs from the U.S. energy and macroeconomic impact were wholesale and retail energy prices for all types of energy, especially natural gas, the energy price and energy consumption forecasts by major end-user categories, along with changes in such variables as gross domestic product, employment, industrial production, inflation, and personal income. Alternative forecasts for California were produced using the Global Insight econometric model for the state, which was revised to explicitly consider the U.S. and California natural gas price levels under the three scenarios, the pattern of energy use by fuel type by major economic sector in California, and the relative economic performance of California as compared with that of the U.S. economy. We combined time-series data on regional and U.S. energy prices and the energy use shares (in Btu) by sector to derive a weighted price of energy by sector in California. We then econometrically estimated the historical relationship between the economic growth of a sector and its weighted average price of energy in order to capture how the changes in energy prices had historically affected levels of employment and output in the sector. The changes made in the California econometric model allowed us to consider the direct effects of the prices of natural gas under the scenarios, plus the prices for coal, oil, and electricity produced during the two prior steps. An appropriate feedback mechanism was created in the model to capture the induced effects of the changes in energy prices on the California economy. Using the national economic impact in the form of drivers from the U.S. macroeconomic models produced the indirect effect under the High- and Low-price scenarios in the California model. Revisions to the Global Insight California Forecast Model We revised our California economic model to directly consider the pattern of energy use by fuel type in the major economic sectors, including manufacturing sectors at the three-digit NAICs sector level. Revised models included sector-specific unit energy cost equations, which were calculated based on energy weights derived from input/output (I/O) coefficients published by the Bureau of Economic Analysis (BEA) and energy use by economic sector in California as reported by EIA. The weights indicate the percentage shares in Btu of each major type of energy (e.g., coal, oil, natural gas, and electricity) used to produce a unit of output for an individual economic sector. Since production functions vary by economic sector, shares of energy use are also different; for example, the chemical industry uses large amounts of natural 28

31 gas as a feedstock and to heat boilers; the primary metals industry uses large amounts of electricity. The modeling framework allowed energy prices to affect unit energy costs in a given sector, considering the relative intensities of various types of energy used in that sector in California. Thus the total expenditures for energy by an economic sector is a function of two factors: 1) the amount of the different types of energy it uses (e.g., coal, natural gas, petroleum, electricity, etc.) in Btu to produce a unit of output, with the percentage distribution of total Btu by energy type representing the energy shares; and 2) the costs of each type of energy in equivalent terms or $/MMBtu. Sector-specific unit energy cost variables directly feed into employment equations, and then flow through the rest of the model in two ways: as employment directly feeds into some equations, and as a feedback into several other concepts. We combined time-series data on regional and U.S. energy prices by economic sector and the energy weights to econometrically estimate the historic relationship between sectoral economic performance (as measured by employment or value added) and its weighted average price of energy. This approach enabled us to enter the forecasts of U.S. and California retail energy prices prepared by our Energy Group into our revised California model to show how the changes in energy prices affect the performance of individual economic sectors, focusing on the three-digit NAICS manufacturing sector and the NAICs supersectors. We also introduced energy prices into other equations in our model to capture the relationship between changing energy prices and household disposable income, especially on how changes in the retail cost of electricity affect disposable income. The derivation of the unit energy cost equations is shown below: Using the following symbols: Natural Gas = n Electricity = e Oil = o Coal = c Unit Energy Cost = EC California = CA United State = US NAICS sectors are represented with "I"; energy prices are represented as P with a subscript identifying the type of energy source and a superscript identifying whether it is intended for use in equations for either the United States or California. As a result, the unit energy cost equation for an i th sector in the U.S. economy can be written as: EC USi = β * P + β * P + β * P + β * P ni US n ci US c oi US o o ei US e And for California: EC CAi = α * P + α * P + α * P + α * P ni CA n ci CA c oi CA o o ei CA e The β s in the above unit energy cost equations represents the amount of a particular energy source (in MMBtu) required to produce one unit of output for the sector. We defined these as the energy weights, and as noted above, they were based on the national I/O tables. The weights were adjusted to reflect patterns of energy use by energy type by economic sectors in 29

32 California; α represents these energy weights for California in the above unit cost equation. In the supply and demand discussion, we noted that manufacturing sectors and electricity generation in California generally have higher shares of their energy inputs on a BTU basis from natural gas than at the U.S. level. We included EC CAi /EC USi as a new independent variable in each employment equation in addition to the other independent variables already used in them. Therefore, the larger differential between U.S. and California natural gas prices, the larger the impact on the California economy. Employment equations are the core of the Global Insight California model, so the effect then flows through other concepts, including GSP, wages, and personal income. Through a feedback mechanism, the effect of changes in other variables flows back to employment, which again flows through rest of the concepts, and so on. This mechanism continues until the net impact of natural gas prices is completely absorbed in the solution. 30

33 ECONOMIC IMPACTS Economic Effects of Changes in Natural Gas Prices Higher natural gas prices have a negative impact on a state economy in three ways: 1) direct effects the amount end users must spend for natural gas rises; 2) indirect effects end users buy fewer inputs from suppliers because their sales decline due to price increases caused by higher natural gas prices; and 3) induced effects employment declines at the directly and indirectly affected firms lead to drops in local expenditures of disposable income. In the short run, a business has two options to offset the higher natural gas costs: pass along some or all of them to its customers, and reduce other non-gas costs by lowering employment or buying less from their suppliers. In the short run, households are also forced to spend more for natural gas and less for other goods and services. Supplying firms therefore experience reduced sales, leading them to also reduce purchases from their suppliers, and so forth. If a manufacturing firm that spends more to buy natural gas furloughs workers or reduces wages, spending at local retail stores will decline, causing them to reduce employment as sales fall, and so on. The dollar impacts presented herein are expressed either in nominal or current dollars (i.e., not adjusted for inflation) and in constant year 2000 dollars (i.e., adjusted for inflation). Nominal dollars are used where appropriate as they measure what goods and services will cost a businesses and households when they are purchased. Households make their spending decisions based on nominal prices, not on an inflation-adjusted basis. Real impacts are the most accurate measures of the level of economic activity because they take price changes into account (i.e., they adjust for inflation). We present real impacts in constant year 2000 dollars because that is the current base year for the real GDP deflator (i.e., in 2000 constant and nominal dollars are the same, or the index = 100), and because real impacts in our U.S. macroeconomic and California models are generated in year 2000 dollars. Impacts on the U.S Economy Using the results of the three energy forecasts, Global Insight prepared three alternative U.S. macroeconomic forecasts based on the alternative price scenarios. The primary purpose of the three macroeconomic forecasts was to generate the values for the exogenous variables (e.g., GDP, industrial production, employment by major sector, industrial production, and inflation) that were used in the enhanced California model. We also used the results of the three energy forecasts as exogenous variables in our enhanced California model, specifically forecasts of retail energy prices for California and the United States and total energy consumption of natural gas and electricity by major end user in California. The U.S. impacts are presented to provide a context for the California analysis in terms of the direction and magnitude of the impacts by variable (e.g., gross domestic products, employment, income, etc.), and the distributional effects (i.e., which economic sectors are likely to be most significantly affected). The U.S. impacts are not directly comparable to the California impacts without an adjustment because the U.S. impacts were based on different wholesale prices for natural gas under the Low and High-price scenarios as reflected in the basis differentials. The percent difference in the Henry Hub prices between the Low and High-price scenarios is less than the percent difference in the Topock Border price; this is shown in Figure 5 where the dashed lines (Henry Hub price) for the High and Low scenarios are closer to each other than the solid lines (Topock Border price) for the same two scenarios. 31

34 The forecast results showed that the impacts of higher natural gas prices on the U.S. economy would generally be lower, in percentage terms, than the impacts in California, though there were some differences. For example, our analysis showed that real GDP under the High-price scenario would be about 0.5% lower in 2016 than under the Low-price scenario, while total employment would be 0.3% lower. In order to correct for the differences in the level of wholesale natural gas prices between the U.S. and California under the three scenarios, we calculated implicit elasticities for key indicators such as GDP and employment based on increases in natural gas prices from the Low-price scenario. For example, the implicit elasticity for real GDP between the Low-price and High-price scenarios in 2016 was -0.02, which means that, on average, for each 1% increase in the price of natural gas between the Low and High scenarios, real GDP would fall 0.02%. The implicit elasticities confirmed that economic impacts of high natural gas prices on the US economy would generally be slightly less than in California; however several differences emerged: The U.S. impacts were more symmetrical, meaning that as natural gas prices rise, the percent declines in economic activity are relatively constant over the entire range of the price change. By contrast, in California the percentage drops in economic activity decline as the price gets higher. This difference is likely due to the following factors: 1) California s natural gas intensive sectors, especially in manufacturing, have already adjusted to higher natural gas prices; 2) California generates a substantially higher share of its electricity using natural gas than does the rest of the United States, so higher electricity prices caused by higher gas prices have an additional adverse impact, and 3) other costs are higher in California (e.g., labor, land, taxes, etc.) so that beyond a certain level, rising natural gas prices have a declining impact. The U.S. manufacturing sector would be affected more significantly than would the California manufacturing sector, as was shown by the higher implicit elasticities between the Low and High scenarios for both U.S. industrial production and U.S. manufacturing employment. The differences in the percent impacts between California and the United States would not be due to differences in economic structure as the composition of the California and U.S. economies are quite similar; for example, California employment shares in the manufacturing and the private, services-producing sectors are currently 10.4% and 67.1% as compared with 10.7% and 67% nationally. Finally, the impacts on real incomes of higher retail natural gas would be somewhat higher in California than nationwide, in part because of the higher share of electricity generated in California from burning natural gas. Impacts on the California Economy Direct Impacts Total economic impacts in California will be generated initially by direct economic impacts, or the amount of money spent by final users to purchase natural gas in each of the three scenarios. In the High-price scenario, the amount of money spent will be greater than that spent in the Middle-price and Low-price scenarios, or a cost increase to households and businesses as compared with spending in the other two scenarios. By contrast, in the Low-price scenario, the amount of money spent to purchase natural gas will be substantially lower than in the other two scenarios, resulting in a cost savings to households and businesses. Table 1 presents expenditures for natural gas under the three scenarios in 2010 and 2016, by major end user, with results presented in both nominal and real terms. 32

35 Table 1: Direct Effects Purchases of Natural Gas by Major End User by Scenario Total Direct Spending for Natural Gas in California by End User by Price Scenario (Billions of nominal $) 2010 Low Middle High Residential $ 5.9 $ 6.8 $ 7.7 Commercial $ 2.5 $ 2.9 $ 3.3 Industrial $ 6.2 $ 7.5 $ 8.6 Electric Generation $ 4.3 $ 5.3 $ 6.3 Total $ 18.8 $ 22.5 $ 26.0 Change from Low $ 3.6 $ Low Middle High Residential $ 6.4 $ 8.1 $ 9.7 Commercial $ 2.7 $ 3.5 $ 4.3 Industrial $ 6.6 $ 8.8 $ 10.8 Electric Generation $ 4.6 $ 6.5 $ 8.3 Total $ 20.2 $ 26.9 $ 33.2 Change from Low $ 6.6 $ 13.0 Total Direct Spending for Natural Gas in California by End User by Price Scenario (Billions of 2000$) 2010 Low Middle High Residential $ 4.7 $ 5.5 $ 6.2 Commercial $ 2.0 $ 2.3 $ 2.7 Industrial $ 5.0 $ 6.0 $ 7.0 Electric Generation $ 3.4 $ 4.3 $ 5.1 Total $ 15.2 $ 18.1 $ 20.9 Change from Low $ 2.9 $ Low Middle High Residential $ 4.4 $ 5.6 $ 6.8 Commercial $ 1.9 $ 2.5 $ 3.0 Industrial $ 4.6 $ 6.1 $ 7.6 Electric Generation $ 3.2 $ 4.6 $ 5.8 Total $ 14.1 $ 18.8 $ 23.2 Change from Low $ 4.6 $ 9.0 These direct effects will spread through the California economy via the indirect and induced effects described previously. Global Insight forecast models produce estimates of the total net change in economic activity, so the direct, indirect, and induced effects are included in the results presented herein. It is important to note that these economic impacts are cumulative and will occur with time as firms and households make dynamic adjustments to natural gas price levels. The economic impacts in 2016 will be the result of adjustments made in by the affected firms and households. Because our models use time-series data, they are able to capture these adjustments over time. Finally, the economic impacts produced by our models include the direct, indirect, and induced effects. 33

36 In the first part of this section, we discuss the total economic impacts on the statewide economy, first presenting the most important indicators of the level of economic activity such as total employment, GSP, and personal income. We then present the economic impacts by major end user groups (e.g., residential, commercial, industrial and electric power generation) for such indicators as expenditures for natural gas and electricity. Finally, we present impacts in Northern and Southern California. Appendix A contains tables presenting the economic impacts in more detail. Appendix B presents the results of our interviews with large natural gas users, and Appendix C presents a more detailed discussion of the employment impacts, focusing on the manufacturing sectors that would be most significantly affected by the natural gas price changes. The analysis below focuses on the differences in the levels of economic activity for two comparisons: 1) between the Middle-price and Low-price scenarios, and 2) between the Highprice and Low-price scenarios in California in 2010 and Global Insight and Advisory Committee did not designate any of the three price scenarios as a Base or Most Likely case; rather our objective was to show how the level of economic activity in California would differ under the three price scenarios. We present both numerical differences and percent differences for the two comparisons. Finally, the output and income impacts are presented, for the most part, in real terms as it is necessary to correct for the effects of energy price levels on the rate of inflation; in our judgment real variables provide the most accurate measure of economic impacts. Employment Table 2 shows that by 2016, total employment in California under the Middle-price scenario would be million jobs, or 97,700 fewer than in the Low scenario, while for the High-price scenario, total employment would be million, a decline of 163,300 jobs from the Lowprice scenario. The table shows that employment impacts increase over time as the absolute and percent differences among the Middle- and Low-price, and the High- and Low-price scenarios, are smaller in 2010 than in This increase in absolute and percent differences over time also occurs for the output and income variables, indicating that the businesses and households directly affected by persistently higher natural gas prices will make continual adjustments over time, which means that the level differences between the scenarios by 2016 are cumulative, and are a direct result of changes made throughout the period of analysis. Table 2 also presents employment impacts for the Manufacturing, Private Services-Providing (PSP, and Construction/Natural Resources/Mining sectors. Approximately two-thirds of the jobs lost within the Middle and High scenarios by 2016 would be in the PSP sectors (e.g., Trade, Transportation and Utilities; Information; Financial Activities; Professional & Business Services; Leisure & Hospitality Services, Education & Health Services, and Other Services) which is unsurprising, as the PSP sectors account for about 67% of total employment in California, a share that will continue to gradually increase over time. The decline in manufacturing employment by 2016 under the Middle- and High-price scenarios would be 15,200 and 32,500 jobs, respectively. Manufacturing employment has been declining in California and throughout the United States for many years, but the rate of loss has accelerated recently as California lost more than 328,800 manufacturing jobs in the last five years. We forecast that manufacturing employment will continue to decline in California and nationwide, regardless of the price of natural gas. As a result, higher natural gas prices 34

37 Table 2: Summary of Economic Impacts in 2010 and 2016 Total Non-Farm (Thousands of Forecast Year Jobs) Manufacturing (Thousands of Jobs) Low Middle High Low Middle High 2010 Level 15, , , , , ,531.5 Numerical Diff. from Low % Diff. from Low -0.4% -0.7% -0.4% -0.9% 2016 Level 17, , , , , ,542.3 Numerical Diff. from Low % Diff. from Low -0.6% -1.0% -1.0% -2.1% Private, Services-Providing Sectors Construction, Natural Resources & Forecast Year (Thousands of Jobs) Mining (Thousands of Jobs) Low Middle High Low Middle High 2010 Level 10, , , , , ,024.2 Numerical Diff. from Low % Diff. from Low -0.4% -0.7% -0.6% -1.1% 2016 Level 11, , , , , ,143.6 Numerical Diff. from Low % Diff. from Low -0.6% -0.9% -0.9% -1.3% Forecast Year Low Middle High Low Middle High 2010 Level $784.7 $780.0 $775.4 $115.3 $114.7 $114.3 Numerical Diff. from Low -$4.7 -$9.2 -$0.5 -$1.0 % Diff. from Low -0.6% -1.2% -0.5% -0.9% 2016 Level $922.0 $914.2 $906.6 $127.5 $126.8 $126.2 Numerical Diff. from Low -$7.8 -$15.3 -$0.6 -$1.3 % Diff. from Low -0.8% -1.7% -0.5% -1.0% Real Personal Income (Billions of Real Disposable Personal Income Forecast Year 2000$) (Billions of 2000$) Low Middle High Low Middle High 2010 Level $1,540.5 $1,532.7 $1,526.4 $1,344.8 $1,339.2 $1,333.7 Numerical Diff. from Low -$7.9 -$14.2 -$5.6 -$11.1 % Diff. from Low -0.5% -0.9% -0.4% -0.8% 2016 Level $1,844.8 $1,834.2 $1,824.0 $1,593.2 $1,584.6 $1,575.8 Numerical Diff. from Low -$10.6 -$20.7 -$8.5 -$17.3 % Diff. from Low -0.6% -1.1% -0.5% -1.1% Forecast Year Employment Impacts Wage and Household Income Impacts Real Wage Disbursements (Billions of 2000$) Real Household Income (2000$) Personal Income Impacts Gross State Product Impacts Real Gross State Product (Billions of 2000$) Real Gross State Product Manufacturing (Billions of 2000$) Low Middle High Low Middle High 2010 Level $1,810.3 $1,801.6 $1,794.5 $ $ $ Numerical Diff. from Low -$8.7 -$15.8 -$1.9 -$4.2 % Diff. from Low -0.5% -0.9% -0.7% -1.5% 2016 Level $2,327.6 $2,310.7 $2,297.2 $395.8 $389.1 $382.8 Numerical Diff. from Low -$17.0 -$30.4 -$6.8 -$13.1 % Diff. from Low -0.7% -1.3% -1.7% -3.3% 35

38 under the Middle and High scenarios will increase the rate of decline in manufacturing employment slightly. The impacts on total employment and employment in the private, services-providing (PSP) sectors in California by 2016 are asymmetrical, meaning that as natural gas prices rise the percent declines in employment become smaller over the entire range of the price change. This indicates that as natural gas prices rise, their marginal effect on the levels of total and PSP employment becomes smaller, most likely because other costs (e.g., total labor compensation, land, taxes, etc.) are high so that beyond a certain level natural gas prices have little impact on the cost of production. The reverse is true in the manufacturing sector where the employment impacts are more symmetrical, which reflects the decline in manufacturing employment that we forecast will continue regardless of the price of natural gas. Appendix C presents a discussion of the derivation and meaning of the implicit employment elasticities that led to the finding of the asymmetry of employment impacts in a number of sectors. Personal Income Table 2 shows that changes in real personal income would be large in absolute terms; by 2016 total real personal income under the Middle-price scenario would be $10.6 billion less than in the Low-price scenario, while the decline under the High-price scenario would be $20.7 billion. The impacts on real wage disbursements are larger in percent terms than those for total personal income. As shown in Table 2 in 2016, real wage income under the High-price scenario would be 1.7% lower than in the Low-price scenario, while the percent fall in real personal income would be only 1.1%. The decline in total wage income is a product of the drop in employment and the accompanying loss in wages per each eliminated job which usually grow at an annual rate of about % in nominal terms. The non-wage components of personal income will be less affected than will wage and salary earnings, and since the latter are about 55% of total personal income, the percent decline in total personal income would be smaller. The impact on real wage disbursements will occur in part because higher natural gas prices will increase inflation, but more because the cumulative, adverse effects of higher gas prices will lower the level of real GSP in California, which in turn would result in declines in employment and wages. We estimate that in the Middle-price and High-price scenarios, per-household real personal income in 2016 will be $643 and $1,301 less than Low-price scenario level of $127,452. These real income decreases are produced by the cumulative, overall declines in economic activity that will have occurred by 2016 as shown in Table 2 (i.e., lower levels of real GSP, employment and income), combined with inflationary effects of higher energy prices. The combination of the decreases in real per household disposable income under the Middle-price and High-price scenarios, along with the higher spending for natural gas and electricity, will adversely affect California s consumers. Households will have less income to spend in the High- and Middleprice scenarios; at the same time they will have to alter their spending patterns by using more of their income to purchase natural gas and electricity, leaving less available to buy other goods and services. Gross State Product Table 2 shows that real GSP, or real value added output, in 2016 would be $30.4 billion (year 2000 dollars) and 1.3% less under the High-price scenario than under the Low-price scenario, while the difference between the Middle- and Low-price scenarios would be $17.0 billion and 36

39 0.7%. In 2010 and 2016, the percentage differences in real GSP under the two comparisons are larger than the percent differences in employment for the reasons noted above. The level of real value added output in manufacturing sector is more sensitive to the price of natural gas than other sectors as it more directly affects production costs, so that the percent differences in real Manufacturing GSP between the Middle- and Low-price, and the High- and Low-price scenarios are significantly higher than they are for difference in total real GSP as is shown in Table 2. We also calculated the changes in nominal (i.e., not adjusted for inflation) GSP as follows: Year 2010: Middle vs. Low-price scenario shows a decline of $12.0 billion; High vs. Lowprice scenario shows a decline of $22.0 billion Year 2016: Middle vs. Low-price scenario shows a decline of -$23.4 billion; High vs. Low-price scenario shows a decline of $42.2 billion The difference in output extends well beyond the simple change in employment. Our simulation shows that by 2016, for every job lost between the Middle- and Low-price scenarios, real GSP in California would fall $173,600, while for every job lost between the High and Low scenarios, real GSP would drop $184,400. This difference would occur because as natural gas prices increase, the value of real GSP continues to decline at about the same rate while the rate of job losses slows (i.e., jobs have already been largely eliminated, so the savings must come through other actions, such as reducing prices). Table 2 shows the relationship between the types of economic impacts that are produced by the direct effects, or the increases in spending for natural gas. First, employment declines as businesses eliminate jobs to reduce costs, or due to lower demand for their goods and services caused by higher prices. At the same time, households have to spend more for natural gas and electricity, leaving them with less to spend for other items. The drop in employment results in a decline in real wage and salary earnings. Since wage and salary earnings are about 56% of total personal income in California, the fall in personal income shown in Table 2 includes the drop in real wage and salary earnings. Next, total value added as measured by GSP (i.e., the amount of the U.S. GDP produced in the California) falls. Since personal income is about 81% of the California GSP, the decline in GSP shown in Table 2 in turn includes the drop in personal income. Stated another way, in Table 2, the drop in wage and salary earnings is less than and included in the decline in personal income, which in turn is less than and included in the fall in GSP. Multipliers One measure of the significance of economic impacts is the economic multiplier, which is the ratio of the total change in the level of a specific variable (often employment) over the direct effect as measured by the same variable. For example, if a new plant starts operating in a region, employs 500 workers, and is later estimated to generated a total increase in regional employment of 1,250 jobs (i.e., the 500 direct jobs plus 750 additional jobs created through the indirect and induced effects), the employment multiplier would be (1,250/500) or 2.5 meaning that for every new direct job, 1.5 additional jobs were created in the region. It should be noted that the total increase in employment results from both the purchases made by the 500 new employees (i.e., the induced effect) and the purchases of goods and services from suppliers located in the region (i.e., the indirect effect). In this study, the direct effect is the change in the value of natural gas purchased by end-users under the three price scenarios presented in Table 1. As was noted above, businesses and households will adjust to higher gas prices over time by taking actions that reduce the amount they will have to spend for natural gas. In theory, the value of the total change in economic 37

40 activity should be greater than the value of the direct effect; therefore, we constructed an implicit multiplier using GSP or real, valued added output, to measure the significance of the economic impacts between the two comparisons (i.e., the Middle-price scenario as compared with the Low-price scenario, and the High- as compared with the Low-price scenario), and also to test the performance of our enhanced California model. The multiplier was the ratio of change in real GSP between two scenarios over the change in the amount spent for natural gas in constant 2000$ between the same two scenarios. We estimated the following multipliers: Middle- vs. Low-price scenario in 2010: an $8.7 billion decline in real GSP/$2.9 billion increase in natural gas purchases, or a multiplier of High- vs. Low-price scenario in 2010: an $15.8 billion decline in real GSP/$5.8 billion increase in natural gas purchases, or a multiplier of Middle- vs. Low-price scenario in 2016: an $17 billion decline in real GSP/$4.6 billion increase in natural gas purchases, or a multiplier of High- vs. Low-price scenario in 2016: an $30.4 billion decline in real GSP/$9 billion increase in natural gas purchases, or a multiplier of Note that, as is consistent with the results presented above, the multiplier increases over time as higher prices persist and as businesses and household continue to adapt to them. Similarly, the multiplier for the High- to Low-price scenario comparison is lower than that for the Middle- to Low-price scenario comparison, showing again that the marginal economic impacts of each additional 1% increase in the price of natural gas decline as the price climbs. We also derived an implicit multiplier that shows the change in real GSP per each job lost under the two comparisons for the two analysis years. Middle- vs. Low-price scenario in 2010: Real GSP decline of $144,200 per job eliminated. High- vs. Low-price scenario in 2010: Real GSP decline of $144,300 per job eliminated. Middle- vs. Low-price scenario in 2016: Real GSP decline of $173,600 per job eliminated. High- vs. Low-price scenario in 2016: Real GSP decline of $186,400 per job eliminated. The two sets of multipliers presented above have consistent results and further confirm that: 1) marginal economic impacts decline in percentage terms as natural gas prices rise; and 2) the negative economic impacts of sustained higher natural gas prices become larger over time in both percentage and absolute terms, as the direct effects of higher natural gas prices spread through the California economy. Impacts on Major End Users This section presents the direct impacts on the major end-users of natural gas: residential, commercial, industrial, and electricity generation. The reactions by the major types of end-users to the alternative natural gas price levels considered in this study will determine the level and characteristics of economic impacts by Households In our linked models, the change in natural gas prices is passed along to both consumers and businesses directly in the form of higher expenditures for natural gas and indirectly through higher electricity costs. California Households comprise one half the Consumer units in the Western Region of the BLS Consumer Expenditure Survey (CES), and the three California 38

41 metropolitan areas surveyed Los Angeles, San Francisco, and San Diego have expenditure patterns near the Western Region average; therefore, we assumed that the Western Region results mirror California to a very similar degree. In 2003, the CES indicated that combined spending for both natural gas and electricity per household was $1,174, or 2.2% of the annual income before taxes of $52,506; consisting of $320 for natural gas (27.2%) and $854 for electricity (72.8%). Since electricity and natural gas spending comprises a small share of CA household purchases, it is unlikely that higher natural gas prices will have a major, adverse impact on household behavior. Even if natural gas prices doubled, although the amount spent for natural gas could also double in the short-run, the far larger share spent for electricity would increase, but not as much; the retail price of electricity would not increase 100%. This isn t to say there wouldn t be painful adjustments: for every additional dollar spent on natural gas and electricity one dollar must be subtracted from elsewhere in the household budget. We estimated the increases in direct spending for natural gas and electricity per household in California under the three price scenarios in 2010 and The results are presented in Table 3. In order to confirm the accuracy of our results, we also estimated the annual spending per California household in 2003 for natural gas and electricity at $361 and $838, respectively, or a total of $1,199 and very close to the Consumer Expenditure Survey (CES) figure of $1,174, thus indicating that our models would yield accurate spending estimates. The figures in Table 3 are presented in nominal dollars since they indicate the actual amount people will have to pay purchase the two types of energy in 2010 and Table 3: Expenditures per Household on Natural Gas and Electricity by Scenario Natural Gas Spending per Household (Nominal $) Low Middle High 2010 Level $440 $510 $578 Numerical Diff. from Low $70 $138 % Diff. from Low 15.9% 31.4% 2016 Level $440 $558 $673 Numerical Diff. from Low $118 $233 % Diff. from Low 26.8% 53.0% Electricity Spending per Household (Nominal $) Low Middle High 2010 Level $937 $970 $1,002 Numerical Diff. from Low $33 $65 % Diff. from Low 3.5% 6.9% 2016 Level $1,097 $1,152 $1,203 Numerical Diff. from Low $55 $106 % Diff. from Low 5.0% 9.7% Average Household Personal Income (Nominal $) Low Middle High 2010 Level $144,024 $143,623 $143,436 Numerical Diff. from Low -$401 -$588 % Diff. from Low -0.3% -0.4% 2016 Level $187,677 $187,208 $185,954 Numerical Diff. from Low -$469 -$1,723 % Diff. from Low -0.2% -0.9% 39

42 By 2016 under the High-price scenario, annual spending per household for natural gas would be $233 higher than under the Low-price scenario, due primarily to the higher Topock border price, which in turn would increase the retail natural gas prices in California. At the same time, annual spending per household for electricity would be $106 higher, for a combined household spending increase of $339. The most significant concern is the effect of the High-price scenario on low-income households (those with annual consumption expenditures of less than $25,000); in 2003 they spent more than 3% on natural gas and electricity and would be more affected by the declines in real disposable household personal income. Low-income households in California would face a more difficult adjustment for two reasons: 1) a large share of their household expenditures are for very essential or non-discretionary items (e.g., food, rent, utilities, transportation, clothing) so reducing purchases of these items to spend more for natural gas and electricity is difficult, and 2) they already spend a higher share of their household income on utilities. According to the 2003 CES report, Southwest region households in the $5,000 20,000 income range spent 3.2% of their income for electricity and natural gas, far more than the 2.4% by the average income household. The three income brackets only represent 22% of the population. Figure 10 shows the growth in the annual combined expenditures for natural gas and electricity by households by scenario in nominal dollars. The combined expenditures were obtained by adding the natural gas and electricity expenditure figures presented in Table 3. For example, under the Low-price scenario by 2016, California households would spend $1,537 (nominal $) for natural gas and electricity as compared with $1,710 in the Middle-price scenario and $1,876 in the High-price scenario. Figure 10: Total Expenditures for Natural Gas & Electricity by Household by Scenario Annual Expenditures per Household $2,000 $1,900 $1,800 $1,700 $1,600 $1,500 $1,400 $1,300 $1,200 $1, Low Price Scenario Middle Price Scenario High Price Scenario As natural gas prices rise, the share of the combined expenditures for electricity declines: in 2016, electricity would account for a 71.3% share under the Low-price scenario, a 67.3% share 40

43 in the Middle-price scenario, and a 64.1% share in the High-price scenario. Electricity accounts for approximately two-thirds of the combined expenditure, so that as electricity prices rise due to higher natural gas prices, consumers will reduce their use of it to lower spending. It is also likely that the percent increase in the residential retail electricity price will be less than the percent increase in the residential retail natural gas price. Table 3 presents average household income levels in nominal dollars. The percent decreases in nominal average household income by 2016 under the Middle- and High-price scenarios shown in Table 3 are less than the corresponding percent declines for real household income shown in Table 2, indicating that the higher nominal energy prices result in higher inflation and eventually in lower real (or inflation-adjusted) levels of economic activity. The increases in spending by households presented in Table 3 were based on the retail natural gas and electricity prices in California forecast by the Global Insight energy model, which in turn were based on the three price scenarios. A potential concern is that the future levels of retail electricity prices in California could be higher than we have forecast because of the use of natural gas by industrial customers who co-generate electric power and sell it to State utilities. The EIA divides the amount of natural gas used by industrial customers for producing steam and co-generating electricity into industrial demand and electric generation demand: the CEC assigns the entire amount of this natural gas use to electric generation demand. The cogeneration issue is significant because a substantial share of the electric power produced in California comes from co-generators. The relationship between the natural gas demand by cogenerators and the effect on electricity prices is discussed in more detail below in the Merchant Power Company section of Appendix B. It is difficult to estimate how much higher retail electricity prices could be under the three price scenarios due to this pass-through; at the same time the higher electricity prices would provide an additional incentive for customers to further reduce their consumption of electricity. For these reasons, we conclude that increases in spending by households for electricity presented in Table 3 are conservative. For example, if by 2016 retail residential electricity prices in the High-price scenario were 10% higher than we have forecast, annual spending for electricity would increase from $1,203 up to as much as $1,323; the actual increase in annual electricity spending would be less than this figure as annual household consumption of electricity would rise less than 10% because of the higher price. Figure 11 shows the consumption of natural gas and electricity in MMBtu per household by scenario. The figure clearly shows that higher retail natural gas and electricity prices would result in California households reducing the amount of energy they use in order to lower spending. Commercial Table 1 presents the changes in purchases of natural gas under the three scenarios for commercial customers. The table shows that under the Middle scenario in 2016, purchases of natural gas (in nominal dollars) by commercial customers would be approximately $3.5 billion, compared to only $2.7 billion under the Low-price scenario, but less than the $4.3 billion spent under the High-price scenario. These direct effects generate the changes in total economic activity presented above in Table 2 and in the tables in Appendix A. Employment in the PSP sectors would not experience the large percentage change in employment as does manufacturing, but the absolute numbers of jobs that will be eliminated would be far higher. Appendix Table A-2 demonstrates that by 2016, employment in the PSP sectors in the Middle-price scenario would account for 67,400 fewer jobs than would employment in the Low-price scenario, or 69% of the total job loss that year. 41

44 Figure 11: Consumption of Natural Gas & and Electricity per Household by Scenario MMBtus/Household/Year Low Price Scenario Middle Price Scenario High Price Scenario Similarly, under the High-price scenario, employment in the private, services-producing sectors in 2016 would be 108,000 fewer jobs than in the Low-price scenario, comprising 66% of the total job loss. Industrial The employment impacts in the manufacturing sector, the economic sector where most of the industrial end users of natural gas are found, were covered above. Table 1 above provides estimates of direct spending for natural gas by industrial customers, and shows that, on average, the amount spent for natural gas by industrial customers is about 2.5 times greater than the amount spent by commercial customers. Under the High-price scenario, industrial customers are projected to spend approximately $10.8 billion (nominal dollars) for natural gas in 2016, well more than the $8.8 billion and $6.6 billion seen in the Middle- and Low-pricescenarios. Government Government employment would be largely unaffected under the three price scenarios. The first two categories barely registered a change in either scenario, only the military which has the option of closing bases and relocating elsewhere was affected. Appendix Table A-2 shows that under the High-price scenario by 2016, total government employment would be only 8,100 fewer jobs than under the Low-price scenario. 42

45 Electric Generation One of the primary direct effects of alternative natural gas prices will be on the cost of generating electricity because California generates a much higher share of its electricity using natural gas than does the rest of the county. According to the EIA, in 2001, just more than 49.0% of electric power generated in California was produced by burning natural gas; only 1.0% was made by burning coal, as compared with U.S. shares of 16.4% and 51.2%, respectively. Since California imports about 20% of its energy, natural gas s share of total electric power used in California is about 41.0% (CEC, 2005 Integrated Energy Policy Report). Higher natural gas prices will directly raise the price of generating electricity, requiring businesses and households to pay more to obtain it and providing them an incentive to use less of it. We estimate that by 2016, total purchases of natural gas by California s electric utilities used in generating electric power under the Middle-price and High-price scenarios will be $6.5 billion and $8.3 billion, substantially higher than the $4.6 billion under the Low-price scenario. The higher prices for natural gas will be reflected in higher retail electricity prices, leading to the negative impacts on households noted above. Impacts in Northern and Southern California At the request of the Advisory Committee, Global Insight estimated the economic impact for Northern and Southern California, with the former generally aligned with Pacific Gas and Electric s natural gas service area as far south as the southern boundary of Tulare County. These two areas are shown in Figure 12, and were based on the current service areas for the major natural gas companies that provide retail service in the state. Southern California was defined as 100% of the nine counties in yellow: Imperial, Los Angeles, San Luis Obispo, Santa Barbara, Orange, San Diego, Ventura, Riverside, and San Bernardino; and portions of the following counties in brown: (based on household coverage): Kern County (Bakersfield) 50%; Tulare (Visalia-Porterville) 95%; Kings (Hanford) 80%; Fresno, 5%. The shares of households served in counties where more than one company provides natural gas services were estimated using information provided by members of the Advisory Committee. The remaining part of the state was defined as Northern California, and consists of all the counties in white and two of three shown in brown. Employment Southern California contains 9.1 million of the 15 million non-farm jobs in California (61% of the total); so it is unsurprising that higher natural gas prices would eliminate more jobs in the lower half of the state than in the less-populated northern half. In comparing the changes between the High- and Low-price scenarios, we forecast that 101,400 of the total statewide decline in employment of 163,300 jobs by 2016 would occur in Southern California, with the remaining 61,900 jobs lost in Northern California. The major reason that employment impacts would be more significant in Southern California is that it has high shares of statewide employment in a number of sectors, especially in the manufacturing, where the impact of the natural gas prices would be the most significant. In other words, Southern California is over-represented in some major natural-gas intensive sectors, and under-represented in others that are much less dependent on natural gas. For most economic sectors, the shares of statewide employment in Southern California range from 60 70%, roughly in line with regional shares of the California population and total employment. We identified nine sectors for which shares of statewide employment exceeded 43

46 Figure 12: Northern and Southern California Study Areas 70%, indicating that they are over-represented in Southern California. All nine of these sectors are in manufacturing, due in large part to the broad manufacturing base of the Los Angeles region, and especially because of the heavy concentration of defense and aerospace firms in southern California. By contrast, some employment sectors are very under-represented in Southern California, indicating that they are likely over-represented in Northern California. We identified eight sectors in Southern California for which shares of statewide employment were less than 60%. Companies in these eight sectors prefer locations in Northern California over those in Southern California. The most obvious of these is beverage manufacturing, as 80% of its statewide employment is in Northern California, which is no surprise: Napa and surrounding counties dominate the wine industry in California. The State Government employment share is also high, as the State Capitol of Sacramento is located in the northern half of the state. Finally, industries related to wood products are concentrated in the northern part of the state, whereas computer firms are concentrated in Silicon Valley. 44